eBPF-Based Instrumentation for Generalisable Diagnosis of Performance Degradation

TL;DR

This paper introduces a set of 16 eBPF-based system metrics for fine-grained, application-agnostic diagnosis of performance degradation in data-intensive applications, enabling detailed root cause analysis.

Contribution

It presents a novel, comprehensive set of eBPF metrics spanning multiple kernel subsystems for diagnosing QoS issues in online data applications.

Findings

Metrics effectively identify causes of performance degradation.

Metrics reveal internal architecture constraints of applications.

Approach works across various workload and contention scenarios.

Abstract

Online Data Intensive applications (e.g. message brokers, ML inference and databases) are core components of the modern internet, providing critical functionalities to connecting services. The load variability and interference they experience are generally the main causes of Quality of Service (QoS) degradation, harming depending applications, and resulting in an impaired end-user experience. Uncovering the cause of QoS degradation requires detailed instrumentation of an application's activity. Existing generalisable approaches utilise readily available system metrics that encode interference in kernel metrics, but unfortunately, these approaches lack the required detail to pinpoint granular causes of performance degradation (e.g., lock, disk and CPU contention). In contrast, this paper explores the use of fine-grained system-level metrics to facilitate an application-agnostic diagnosis of QoS degradation. To this end, we introduce and implement $16$ $\textit{eBPF-based metrics}$ spanning over six kernel subsystems, which capture statistics over kernel events that often highlight obstacles impeding an application's progress. We demonstrate the use of our $\textit{eBPF-based metrics}$ through extensive experiments containing a representative set of online data-intensive applications. Results show that the implemented metrics can deconstruct performance degradation when applications face variable workload patterns and common resource contention scenarios, while also revealing applications' internal architecture constraints.